Skip to content
Sill Docs

The /agent page

Some AI assistants read web pages well but struggle to fetch raw JSON files. The /agent page gives them the same information as your signed agent card as an ordinary HTML page: your store, its discovery endpoints and the skills agents can use. Sill publishes one for every verified site.

Where it is served

Section titled “Where it is served”
GET https://edge.sill.so/v1/agent/{site_key_hash}

With a Cloudflare Worker or WordPress plugin install, it is also served on your own domain at https://your-domain/agent, and every HTML page carries a small hidden link to it. Your llms.txt links to it too. See What each install covers.

What the page shows

Section titled “What the page shows”
  • Your store name, with a short description when you have set one.
  • Website and Store policies: your homepage, plus the returns, shipping, currency, price range and policy links you confirmed, when set.
  • Discovery endpoints: your signed agent card, your signed ARD catalog, your MCP server, your llms.txt, your UCP business profile, your /.well-known/api-catalog address, and the public JWKS for checking signatures. The catalog and UCP links appear for stores only.
  • Available capabilities: your live skills, each with its id and a one-line description. These are the same skills your agent card, MCP server and catalog list.
  • How to buy (stores only), Authentication and Verification: two or three sentences each on signed orders and signature checks.
  • A <script type="application/ld+json"> block with the same endpoints and skills as structured data. This block is not signed; check signatures on the agent card itself.

The page has no styling or scripts beyond that data block, so it reads the same in any browser or text-only fetcher.

The /agent page for a demo coffee store listing store policies, discovery endpoints and four skills What an assistant sees when it opens your /agent page.

The page is cached for 60 seconds and may be served from cache for a few minutes more while it refreshes, so skill changes show up within a few minutes. A site whose domain is not verified yet returns 404.

Add it to your sitemap

Section titled “Add it to your sitemap”

Add /agent to your sitemap.xml so search crawlers, including the indexes that AI assistants search, find it:

<url>
  <loc>https://your-domain/agent</loc>
</url>

This only helps when the page is served on your own domain (Cloudflare Worker or WordPress plugin install).

See also

Section titled “See also”